I wish to train an algorithm to detect defects on images of labels. These may be such things as scratches, tears and voids.

I would like to try to train a YOLO algorithm to do this, but it is very computationally intensive and requires all training images to have the bounding boxes identified.

Has YOLO been used successfully to detect defects?

  • $\begingroup$ Depending on the details of the problem there may be an easier method. Are the images taken from a consistent orientation? If they are or if transforming them to be the same is not difficult, then a metric based on a simple image difference between the expected/ideal label image and the image to be processed could be pretty effective. $\endgroup$ – Doug7 Nov 4 '18 at 18:32
  • $\begingroup$ Unfortunately, the images are not taken from a consistent orientation; the images are on bottles, which rotate to random angles in front of the camera(s). $\endgroup$ – Sotades Nov 5 '18 at 7:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.